The Satellite Digital Twin (SDT) provides model-based data monitoring, engineering analysis, conditional maintenance, and high-fidelity simulation in operations and maintenance. It exceeds current machine learning applications for monitoring satellite health and safety by broadening the scope of satellite telemetry, commands, and orbital event data. The timed finite state machine (TSFM) for satellite operations involving operational events leverages this expanded data scope to form a state equation associating satellite states with operational events in telemetry data and event triggers, such as satellite commands and orbital events. The state equation establishes a framework for more proactive and dynamic monitoring and model-based high-fidelity simulations. The model recalibration in an SDT requires a new architecture to learn state profiles from telemetry data and link these profiles with event triggers to establish the state equation.